Using natural language processing (NLP)-inspired molecular embedding approach to predict Hansen solubility parameters
Jiayun Pang, Alexander W. R. Pine, Abdulai Sulemana
Abstract
Hansen solubility parameters can be predicted with good accuracy using only the SMILES of molecules and a BERT deep learning model with finetuning.
Topics & Concepts
Natural language processingArtificial intelligenceEmbeddingComputer scienceComputational Drug Discovery MethodsCrystallization and Solubility StudiesMachine Learning in Materials Science